(459i) Smart Agriculture: Early Detection of Plant Diseases Using Multifunctional Wearable Sensor Data and Machine Learning

The United Nations has estimated that food productivity needs to increase by approximately 60% by 2050 to sustain a projected population of 10 billion. However, plant diseases are a significant contributing factor to global crop loss, ranging from 20 to 40% annually. This adversely affects food production, biodiversity, socioeconomic factors, control costs, and overall global food security. Wearable plant sensors offer promising solutions for smart agriculture. Our research introduces a lower leaf surface-attached multimodal wearable sensor capable of continuously monitoring plant physiology by tracking both biochemical and biophysical signals of the plant and its microenvironment. Integrated into a single platform are sensors for detecting volatile organic compounds (VOCs), temperature, and humidity. Additionally, we have developed an unsupervised machine learning (ML) model to analyze multichannel sensor data for quantitative detection of the tomato spotted wilt virus (TSWV) as early as 4 days after inoculation. The model also assesses various sensor combinations for early disease detection and indicates that a minimum of three sensors, including VOC sensors, are necessary for accurate prediction. By combining advanced sensing technologies with sophisticated machine learning techniques, we pave the way for transformative advancements in smart agriculture and the sustainable management of global food systems.